bioreactor monitoring & control - mti-ca.com · bioreactor monitoring & control ... • sh...
TRANSCRIPT
Bioreactor Monitoring & ControlBioreactor Monitoring & Control
Bioreactor Monitoring & ControlBioreactor Monitoring & Control
• Basic principles of process controlBasic principles of process control• Fermentation monitoring
Di l d– Dissolved oxygen – pH– Temperature– Offgas monitoring– Substrate (glucose)
Useful referencesUseful references• Ogunnaike, B.A., and Ray, W.H., Process Dynamics, g y yModeling, and Control, Oxford University Press, 1994.
• Vogel H C Todaro C L eds Fermentation &Vogel, H.C., Todaro, C.L., eds., Fermentation & Biochemical Engineering Handbook 2nd Ed., Elsevier 1996.Sh l M L d K i F Bi E i i• Shuler, M.L. and Kargi, F., Bioprocess Engineering: Basic Concepts 2nd Ed., Prentice Hall, 2001.
• Van Impe, J.F.M., Vanrollegham P.A.,and Iserentant, p , , g , ,D.I., eds., Advanced Instrumentation, Data Interpretation, and Control of Biotechnological Processes, Kluwer Academic Publishers, 1998.Processes, Kluwer Academic Publishers, 1998.
Process Control Basics: The concept of lprocess control
F
offgasFeed in
Fi
Fgas
Presentation of problem:
1. Control level of liquid
Fo
2. Control concentration of O2 in fermentorLevel probe
∞∞ Broth out
Definitions:DO probe
Air in
Chemostat
FairInput Variables are those that independently stimulate the system and can thereby induce change in the internal conditions of the process
Output Variables are those by which one obtains information about the internal state of the system
Process Control Basics: The concept of lprocess control
F
offgasFeed in
Fi
FgasTwo types of input variables
Those we can manipulate are control variables
Fo
control variables
Those we cannot control are disturbance variables
W t l id tif
Level probe
∞∞ Broth out
We can separately identify variables that we can measure.
DO probe
Air in
Chemostat
Fair
Feedback Control ConfigurationDISTURBANCE
Input Output
Final Control El t
Process
SensorElement
Transmitter
DECISIONTransmitter
INFORMATIONDECISIONController
Control action is taken based on information that there is a process upsetControl action is taken based on information that there is a process upset
Upset occurs when the process output variable deviates from the setpoint.
Feedback Control: Regulatory Control5
Feedback Control: Regulatory Control
3.5
4
4.5 underdamped
critically damped
onse
2.5
3
3 5
overdamped
cess
resp
o
1.5
2
Proc
0
0.5
1
SPdisturbance
0 5 10 15 20 25
time
Feedback Control: Regulatory Control
6
Feedback Control: Regulatory Control
4
5
nse
3
4
ss re
spon
1
2
underdampedcritically dampedChange SP
Proc
es
00 2 4 6 8 10 12 14 16
time
Standard Feedback ControllersStandard Feedback Controllers• Proportional Controller (P)• Proportional + Integral Controller (PI)Proportional + Integral Controller (PI)• Proportional + Integral + Derivative Controller (PID)
offgasFeed in Bioreactor systems generally come set up so that you don’t need to know which type ofneed to know which type of control strategy has been employed – they just do it…
∞∞ samples
Standard:
• temperature
Air in • pH
• dissolved oxygen
Process Monitoring: Hardware SensorsProcess Monitoring: Hardware Sensors
Hardware sensorsHardware sensors– Dissolved oxygen probe – pH probepH probe– Temperature probeOffgas monitoring (IR + flow sensor; mass spec)– Offgas monitoring (IR + flow sensor; mass spec)
– Dissolved CO2 probeSubstrate (glucose) flow sensor– Substrate (glucose) flow sensor
– Fermentor volume (via load cell)
Bioreactor Control: Nonlinear ProcessBioreactor Control: Nonlinear Process
• PI or PID control strategies can be used forPI or PID control strategies can be used for simply managing process variables such as temperature and pressure.
• pH is somewhat more sophisticated to avoid oscillation and offset
• DO is managed by cascade control• Critical variables, such as nutrient feed rate(s)Critical variables, such as nutrient feed rate(s) may not have simple hardware sensors available (glucose monitoring)
We will begin with oxygeng yg• Oxygen is needed for respiration for aerobic cultures.
• Strict aerobes will die without sufficient O2
• Facultative aerobes will not die in the absence of O2 but will produce byproducts due to the need to transfer electrons to metabolites when O2 is not present to do the job.
d b d l d b• Oxygen is monitored by a dissolved oxygen probe
• Oxygen is fed as a gas but can only be utilized in the liquid phasephase
The E. coli we use in the 194 lab is a facultative aerobe and therefore grows best with oxygen but can grow also without it. In the absence of O2,best with oxygen but can grow also without it. In the absence of O2, however, it will produce byproducts that inhibit the growth. Thus, we aim to avoid an oxygen-limited situation in the process.
Challenges with DO controlChallenges with DO control
• Oxygen will dissolve in water up to yg p[DO]sat'd ~ 7 ppm (= 7 mg/L) at 30ºC.
• This concentration is sufficient for bugs to grow, what is limiting is the g g grate of transfer from the air bubble to the liquid.
T 0C x (mol O2/mol H2O) [O2] mg O2/l H2O10 0 0000064 11 37710 0.0000064 11.37720 0.0000052 9.27730 0.0000044 7.83240 0.0000039 6.95450 0.0000036 6.327
Oxygen transfer engineering styleOxygen transfer engineering styleIf we take a snapshot of a bubble interface during abubble interface during a fermentation run, here is what we see
DistanceAir bubble
liquidbubble
Bubble interface
Assumptions: air bubble and liquid are two well mixed systems
Oxygen transfer ‐ engineering style
The oxygen concentrations vary both with distance and with time.
The actual mechanism of oxygen transfer is complex, so engineers use an empirical method to describe the process.
O f (OTR) k (C* C )Oxygen transfer rate (OTR) = kLa(C*‐CL)• C* is the saturation concentration of O2 in the liquid
• C is the actual liquid concentration (measured by DO probe)• CL is the actual liquid concentration (measured by DO probe)
• “kLa” is the “Liquid side mass transfer coefficient”.
Transferring the O2 from the bubble interface into the liquid is the rate‐limiting step to get the air into the liquid.
Oxygen supply vs. demand
Oxygen transfer rate (OTR) = k a(C*‐C )
Oxygen supply is what is transferred from the bubble into the liquid
Omg 2
mM
orOxygen transfer rate (OTR) = kLa(C ‐CL)
If we have a bacteria culture that consumes a lot of oxygen, then the measured concentration of O2 in the liquid can drop to zero.
hl
h
or
2
How can we increase the oxygen supply??
DO
Oxygen supply vs. demand
What cells consume ‐ or the demand
The rate of O2 consumption is the oxygen uptake rate. 2 p yg p
The units are the same as for OTR =>
hlOmg 2
hmM
or
dOOUR is a rate OUR =
dtdO2 mols O2 consumed/time
if O2 is the limiting substrate, and the O2 limitation comes into play, then 2 g , 2 p y,
OTR = OUR
or the demand = supply ==> can the demand be greater than the supply?
The amount of oxygen required by cells depends on their growth rate and on the total number of cells.
Even cells that are not growing need oxygen, and this is called theEven cells that are not growing need oxygen, and this is called the maintenance requirement.
Oxygen supply vs. demand
How can we increase supply??
Oxygen transfer rate (OTR) = k a(C*‐C )
We can either increase kLa or we can increase C*
Oxygen transfer rate (OTR) = kLa(C ‐CL)
How can we get oxygen to move across the bubble interface faster?
40)( 79% N2
21% O2
6.04.0)(1.0 gsL M
wPak
Air bubble
M is the mass of the fluid.
0.1 is experimentally determined (should be determined for each culture broth).
gsis superficial gas velocity.
Increase O2 Supply to BioreactorStrategy 1: Increase flow of gas into bioreactor
Increase O2 Supply to Bioreactor
40 6.0
4.0)(1.0 gsL MwPak
Oxygen controllerFlowmeter
Dissolved oxygen (DO)
Air compressor
oxygen (DO) probeSparger
Increase O2 Supply to BioreactorMixing
Strategy 2: Increase mixer d
Increase O2 Supply to Bioreactor
Motor controller
speed.
4.0)(P 6.0)(1.0 gsL MwPak
Increase O2 Supply to BioreactorIncrease O2 Supply to BioreactorHow can we increase C*?
Oxygen transfer rate (OTR) = kLa(C*‐CL)
C* is the saturation concentration of O2 in the liquid phase…
We can increase it by increasing the concentration of O2 in the GAS phasep ase
Feed pure O2 or a mixture of O2 + air
If this strategy can be avoided, it should be
dd d t t h ifi d O• added cost to purchase purified O2
• safety issues associated with handling pure O2
Increase O Supply to BioreactorIncrease O2 Supply to BioreactorDO Control involves a combination of both -- called Cascade ControlDO Control involves a combination of both -- called Cascade Control
• increasing stir speed
• increasing air flow rate
• There is a maximum possible stir rate
• Air flow rate cannot be increased indefinitely• Air flow rate cannot be increased indefinitely
Flooding is the phenomena where there is so much air blown into the reactor that you create a pocket of air in the middle of y pthe reactor around the impellers – just beating the air and not the liquid
When flooding occurs there is a notable drop in the power supply to theWhen flooding occurs, there is a notable drop in the power supply to the mixer.
pHpH
What is pHWhat is pHPure water dissociates to yield 10-7 moles/L of H+ at 250C:
H2O <----> H+ + OH-
Since water dissociates to produce one OH- ion for each H+ ion, it is obvious thatobvious that
10-7 OH- ions are produced simultaneously.
The product of [H+] and [OH-] always remains constant. When the value for one of the species changes the other also changes accordingly.
[H+] x [OH-] = 10-14[H ] x [OH ] 10
The concentration of H+ ions can be increased when compounds are added which release H+ ions such as H2SO4:
H2SO4 <----> 2H+ + SO42-
Control of pH in BioreactorControl of pH in Bioreactor
Calculate the concentration of hydrogen ions from the pH
[H+] (molar concentration) = 10-pH
eg for pH = 7.0 and 7.2
[H+] = 10-7 molar or 0 0000001[H+] 10 molar or 0.0000001
[H+] = 10-7.2 molar or 0.0000000631
ti th t i i th H b 0 2 d th [H ] b 37%notice that increasing the pH by 0.2 decreases the [H+] by 37%
Control of pH in BioreactorMost bacteria can grow over a wide range of pH, although many enzymes upon which microbial growth depends function only within a narrower range of pH.
Control of pH in Bioreactor
The bacteria then must maintain their internal pH near a fixed optimal value.
Bacteria (E. coli) that grow at neutral pH (6.0-8.0) are called neutrophiles.Bacteria (E. coli) that grow at neutral pH (6.0 8.0) are called neutrophiles.
Regardless of the external pH, the internal pH is maintained at ~7.6.
H i i t i d b i th b f th b t ipH is maintained by ion pumps on the membrane of the bacteria. • Operation of the pumps requires energy input • Effort put into maintaining the pH will be at the expense of other cellular functions • Bugs tend to grow more slowly when the pH is not at the optimum.
Any processes that involve interaction with the external di h t k f t i t ti f t i tmedium, such as uptake of nutrients, secretion of proteins, etc.
will be directly affected by the external pH.
H t lpH control
pH controllerAcid pump p co t o e
pH probe
Acid pump
BBase pump
What causes the pH to change?What causes the pH to change?
• Overfeeding substrate can cause the cells to produceOverfeeding substrate can cause the cells to produce organic acids, such as acetate – pH drops
• Lack of carbohydrate substrate causes the cells to yconsume protein in the media – producing NH3NH4OH – pH rises
• When producing a protein product, cells consume ammonia from the media from the cellular demand f h l ffor more nitrogen – causes the release of a proton –pH drops.
pH control in CHE 194 processpH control in CHE 194 process
• We use NH4OH for base to control the pHWe use NH4OH for base to control the pH. • We don’t need any acid because the process control is designed to avoid overshootingcontrol is designed to avoid overshooting.
• We wait for a pH spike (production of base) d h h l f dand then we turn on the glucose feed pump.
A very slow glucose feed is used to enable h f b i d idgrowth of bacteria and avoid oxygen
limitation. (explained later in this i )presentation)
Temperature ControlTemperature Control
Heat management is another engineering task
Q Q QQ
The amount of heat (Q) to be added or removed depends on the density of the culture, the volume of the culture, and the growth rate of the culture.
HLGR XYVQ
Heat management in a laboratory bioreactor is analogous to that at a large scale
Flowmeter
T tTemperature controllerCooling water
Temperature probe
Heating jacket
At large scale, heating jacket is replaced with a steam jacket
Heating Jacket on Fermentor
Our (lab 109) heating jacket is electrical, not using hot water orOur (lab 109) heating jacket is electrical, not using hot water or steam. We use it to heat the bioreactor and to help maintain temperature. We use cooling water in a little heat exchanger to cool the system. The combination of heating and cooling inputs carefully maintains the temperat remaintains the temperature.
Metabolic Heat Generation
40‐50% of energy produced by substrate catabolism is converted to ATP, the rest is released as heat. Heat of
Glucose + NH3 + O2 ‐‐‐‐‐‐‐> CO2 + H2O + cells
Heat of Combustion (ΔHC) of cells 20 ‐ 25 kJ/kg
YH is the metabolic heat evolved per gram of cell produced
YH depends on the degree of oxidation of the substrate
YH ~ 2.4 kcal/g on glucose
YH ~ 5.6 kcal/g on ethanol
H p g
Total heat evolved (QGR) depends on
XYVQ k l/hYH ~ 8.3 kcal/g on methanol
YH ~ 16.4 kcal/g on CH4Oxygen uptake
rate
HLGR XYVQ Units kcal/hr
For aerobic fermentations,2
12.0 OGR QQ Units kcal/lit∙hr
Where is the heat production the greatest?
250.000
150.000
200.000
100.000
Dissolved Oxygen
0 000
50.000
Dissolved OxygenOptical Density
0.0000.000 20.000 40.000 60.000 80.000 100.000 120.000
HLGR XYVQ Time from inoculation (hrs)
Cooling water additionCooling water additionSGR QQQ
Total heat to be removed by cooling water is due to metabolic heat generation plus heat added from the stirring
TUAQ TUAQ CHeat is removed by running cooling water through the cooling coils in the reactorcoils in the reactor
U is the overall heat transfer coefficient of the coils
AC is the total area of the coils in contact with the broth
ΔT is the “log-mean temperature difference” between the broth and the cooling water
i TTTT outin
outin
TTTTTTTTT
/ln
Offgas AnalysisOffgas Analysis
Offgas composition is an indicator of culture activityOffgas composition is an indicator of culture activity
Bug: glucose + NH3 + 2O2 C4H7O2N (bug) + 4H2O + 2CO2
Respiration:
glucose + 36 Pi + 36 ADP + 6 O2 6 CO2 + 6 H2O + 36 ATP
B d thi i l d i ti f th h i tBased on this simple description of the chemistry, we can see that production of CO2 is an indicator of living cells. A high rate of CO2 production coupled to O2 consumption is 2 2an indicator of growth.
Offgas analysis Offgas analyzeranalyzer
Condenser
Flowmeter
Air out is <79% N2 and <21% O2 and 0.033% <
CO < 6%CO2< 6%
Air compressorAir in is 79% N2and 21% O2 and 0.033% CO2
Sparger
2
Offgas composition is an indicator of culture activityOffgas composition is an indicator of culture activity
Carbon dioxide Evolution Rate (CER) is the rate of production of CO2.
CO2,out – CO2,in Is the CERTime interval
Is the CER
Since we know the concentration of CO2 in the air is 0.033%, bySince we know the concentration of CO2 in the air is 0.033%, by measuring the CO2,out we can determine CER.
Note that this is measured for the whole culture volume soNote that this is measured for the whole culture volume, so we can divide out the culture volume for a more meaningful value.
Offgas composition is an indicator of culture activityOffgas composition is an indicator of culture activity
O – OTime intervalO2,out – O2,in Is the OURLikewise,
We can measure the CER and OUR on‐line using a mass spectrometer.
The mass spectrometer measures concentrations, and we need to know total amounts.
By comparing the O2 and CO2 concentrations with the N2 we can determine what the flow rate out is, since N2 is neither produced or consumed by the bugs.
Calculation of OUR & CERCalculation of OUR & CER
• Handout – example problemHandout example problem
Best practice: design your feeding strategy such that O2 is not the limiting nutrientthat O2 is not the limiting nutrient…Depending on your product, you may choose Carbon or Nitrogen to be your limiting nutrient, and you can choose different limiting nutrients during different phases of the run.during different phases of the run.
When the cells are growing, the feed rate can be determined from:
)(tXY
SX
FS =
FS is the substrate feed rate,
μ is the specific growth rate of the organisms
YX/S is the yield coefficient (rate of cell production/rate of substrate uptake)YX/S is the yield coefficient (rate of cell production/rate of substrate uptake)
X(t) is the time dependent concentration of organisms in the bioreactor.
Can we estimate the cell growth rate and cell concentration from on-line measurements?
Best practice: design your feeding strategy such that O2 is not the limiting nutrient cont’d
Since we know (we can measure) the yield of cells on oxygen, we can estimate the ll t ti b
that O2 is not the limiting nutrient, cont d
cell concentration by
)()( tOURYdt
tdXO
XOUR can be
measured on-line dt O
Together with an estimate of X(t) we can estimate μ from…
dtdX
tX )(1
tdOUR
tXtOURt
0 )()(
)()(IOUR b dttOUR
Yt
OX
0
0 )()( IOUR - can be
measured on-line
Best practice: design your feeding strategy such that O2 is not the limiting nutrient cont’dthat O2 is not the limiting nutrient, cont d
Putting everything back into our Feed rate calculation, we have
)(tXY
SX
FS =
)()()(1
0tXIOURYXtOURF XS )()( 0
0 IOURY
tXY OX
OX
SX
S
If the cells have everything they need (balanced growth), the μ you are estimating may be close to μmax. If your feed rate should exceed the amount needed for the growth of the cells, you will accumulate substrate and it will result in the formation of byproductssubstrate and it will result in the formation of byproducts. It solves the problem to choose some fraction of μmax, such as 0.7μmax
Off-line analysis methods used
for fermentation
Off‐line analysisOff line analysis• Advantages:
– No interfacing required– Flexibility– Low cost– Small sample volumes
• Disadvantages:– Requires removing samplesq g p– Requires tracking the sampling time– Requires operator interaction, potential for biasRequires operator interaction, potential for bias error from operator
Offline Analysis MethodsOffline Analysis MethodsGlucose
• Glucose concentration is measured by a glucose analyzer. For E. coli f t ti l h ld b k t b b t 5 /lit d b lfermentations, glucose should be kept above about 5 g/liter and below about 20 g/liter for optimum growth.
• The Yellow Springs Instruments (YSI) Bioanalyzer is a standard instrument for fermentation analysis It can measure other sugars ethanolinstrument for fermentation analysis. It can measure other sugars, ethanol and methanol, as well.
• A small sample from the fermentor is centrifuged to settle the bacteria and the supernatant is fed through a sipper tube into the analyzer It isand the supernatant is fed through a sipper tube into the analyzer. It is automatically calibrated.
High Performance Liquid Chromatography (HPLC)
• HPLC is useful for measuring fermentation by-products, such as ethanol, organic acids – acetic acid, succinic acid, lactate, etc., and specific amino acids and virtually all small molecules and proteins.
• Each molecule should have a specific method to be analyzed.
Offline Analysis MethodsOffline Analysis Methods
Determining cell mass concentration
Di t th dDirect methods:
•Dry Weight
→ Solids-free medium.
•Packed Cell Volume
•Optical Density Method
→ Spectrophotometer.
→ 550 or 600 nm wavelengths.
→ light absorbance cell mass/volume→ light absorbance cell mass/volume
→ OD ≤ 0.3 → need to measure dry-weight of the cells.
Typical Off‐line analysesTypical Off line analyses
• Optical density (OD) analysis of cellOptical density (OD) analysis of cell concentration
• Glucose can be measured off line• Glucose can be measured off‐line• NH3 can be measured by chemical analysis• Phosphate and other macro‐nutrients• Protein activity assaysy y• HPLC – everything else…